top 10 iot technologies for 2017 and 2018 -...

G00296351

Top 10 IoT Technologies for 2017 and 2018Published: 22 January 2016

Analyst(s): Nick Jones

This research discusses 10 technologies that will be vital for organizations tounlock the full potential of the IoT as part of their digital business strategies.

Key Findings■ The Internet of Things (IoT) demands a wide range of new technologies and skills that many

organizations have yet to master.

■ A recurring theme in the IoT space is the immaturity of technologies and services and of thevendors providing them. Architecting for this immaturity and managing the risk it creates will bea key challenge for organizations exploiting the IoT.

■ In many technology areas, lack of skills will also pose significant challenges.

Recommendations■ Use techniques such as layering and modularity to architect for change so that future

developments in technologies or vendors can be accommodated. Modularity may extend tohardware as well as software.

■ Evaluate IoT technologies using Gartner Hype Cycles and Market Guides to assess theirmaturity and risk and to create roadmaps for IoT solutions to transition from tactical to morestrategic technologies.

■ Catalog the IoT skills gaps in your organization, and train staff or find appropriate partners toaddress them.

Table of Contents

Analysis..................................................................................................................................................2

IoT Security...................................................................................................................................... 2

IoT Analytics..................................................................................................................................... 3

IoT Device (Thing) Management........................................................................................................4

Low-Power, Short-Range IoT Networks............................................................................................5

IoT Processors................................................................................................................................. 6

IoT Operating Systems..................................................................................................................... 6

Low-Power Wide-Area Networks......................................................................................................7

Event Stream Processing..................................................................................................................8

IoT Platforms.................................................................................................................................... 9

IoT Standards and Ecosystems........................................................................................................ 9

Gartner Recommended Reading.......................................................................................................... 10

List of Tables

Table 1. Selected IoT Operating Systems............................................................................................... 7

AnalysisFor many organizations, the IoT will be a cornerstone of their digital business strategies, but it willalso be very disruptive, requiring them to master many new technologies and capabilities. Thetechnologies and principles of IoT will have a very broad impact on organizations. They will affectbusiness strategy, risk management and a wide range of technical areas such as architecture andnetwork design.

This research discusses 10 IoT technologies that are likely to be on every organization's radar.These are certainly not the only important IoT technologies — Gartner advises organizations toconsult its IoT Hype Cycles for details of many more — but these were selected on the basis of theirimportance to a wide range of organizations and IoT solutions. This note focuses on thosetechnologies that are specific to the IoT; many IoT solutions will also use a wide range ofconventional IT technologies that are not discussed here.

The IoT is a very immature domain where product and technology categories aren't yet clearlyestablished, so some of the topics discussed here are less specific technologies than keytechnology areas that can't yet be satisfied by any single product or vendor. In several cases, it'slikely that researchers will develop new technologies and solutions that don't yet exist. One of therecurring themes in these technology descriptions is risk and immaturity. Organizations needingsolutions in the short term can't afford to wait until the IoT is mature, so managing vendor andtechnology risk will be vital to successful IoT deployments. Key principles will include architectingfor change — for example, modularizing designs so that software and even hardware technologiescan be replaced when superior options emerge.

IoT Security

What is it, and why is it important? The IoT introduces a wide range of new security risks andchallenges to the IoT devices themselves, their platforms and operating systems, theircommunications, and even the systems to which they're connected (such as using IoT devices as

of 12 Gartner, Inc. | G00296351

an attack channel). Creating smarter products often removes traditional architectural barriers; forexample, critical car systems are now exposed to mobile apps (see Note 1 for selected examples ofrecent IoT security flaws). Security technologies will be required to protect IoT devices andplatforms from both information attacks and physical tampering, to encrypt their communications,and to address new challenges such as impersonating "things" or denial-of-sleep attacks that drainbatteries. IoT security will be complicated by the fact that many "things" use simple processors andoperating systems that may not support sophisticated security approaches.

When? Many security tools and technologies are available in 2016 — although seldom from onesingle vendor — and often, many aren't well-adapted to current IoT use cases. It's likely thathardware and software advances will make IoT security a fast-evolving area through 2021. Also,some "things" may be long-lived, allowing attackers many years to find vulnerabilities, so securitystrategies and technologies must be flexible and able to evolve as new threats emerge during aproduct's lifetime. New approaches and technologies will emerge to provide new types of solutions(see Note 2).

Who will be impacted? Any organization delivering or using the IoT. All roles involved in deliveringIoT solutions, including chief information security officers (CISOs), architects, designers andprogrammers, must become familiar with IoT risks, security principles and technologies.Organizations must review the security used in the IoT systems they purchase as well as those theydevelop internally.

Cautions: Experienced IoT security specialists are scarce, and in 2016, security solutions arefragmented and involve multiple vendors. New threats will emerge through 2021 as hackers findnew ways to attack IoT devices and protocols, so long-lived things may need updatable hardwareand software to adapt during their life span. Technology cannot address all security issues; manyIoT security problems are related to lack of knowledge resulting in poor design and implementation.Often, the weakest link in IoT security will be people.

IoT Analytics

What is it, and why is it important? IoT business models will exploit the information collected by"things" in many ways — for example, to understand customer behavior, to deliver services, toimprove products, and to identify and intercept business moments. However, IoT demands newanalytic approaches:

■ As we evolve to a future with tens of billions of connected "things," the volume of data to beanalyzed will increase dramatically. This creates problems of scale that will be partly addressedby new platforms (see the Event Stream Processing section). But it also means that there willnot be sufficient staffing or time for human data scientists to analyze the information. This willdemand technologies such as machine learning to identify patterns, which is an area wherevendors such as Google and IBM have made major investments.

■ The traditional IT approach to data analytics has been to collect and store data, then analyze it.However, in an IoT context, this is not always desirable or practical, so new analyticsarchitectures are emerging. First, there may be too much data to store, so analysis of data

Gartner, Inc. | G00296351 of 12

streams must be conducted on the fly. Second, data filtering and analysis may sometimes bedistributed in gateways at the edge of the network or in the "things" themselves to minimizecommunication over slow networks or when communications have undesirable side effects,such as increasing battery consumption in sensor nodes.

■ The data collected from "things" may involve new data types and analysis algorithms. Timeseries data is very common, demanding filters and Fourier transforms. A growing range of"things" are location-aware, demanding geographic information processing.

IoT analytics therefore needs new tools such as high-volume event stream platforms, the ability tooperate on new data types, new technologies such as machine learning, and new architectureswhere analytics is distributed throughout the network of things.

When? New analytic tools and algorithms are needed now, but as data volumes increase through2021, the needs of the IoT may diverge further from traditional analytics.

Who will be impacted? Data scientists, business intelligence staff, and statisticians performingbusiness analytics will be impacted. Business strategists must be able to exploit new insights andreact more rapidly to insights. Network designers must plan for new traffic patterns.

Cautions: Traditional business intelligence and analytics staff may lack skills in areas such asstreaming analytics and time series data. As IoT devices proliferate, analytics will be able togenerate very personal insights (for example, from monitoring the smart home), so data privacy andacceptable use will become major challenges. Analytics will pose business challenges around dataaccess and ownership; for example, is the data from a ship's engine the property of the ship owneror the engine manufacturer?

IoT Device (Thing) Management

What is it, and why is it important? Long-lived nontrivial "things" will require management andmonitoring. This includes device monitoring (for example, are devices still alive, are they connected,and what is their battery status?), firmware and software updates, diagnostics, crash analysis andreporting, physical management (for example, installation, retirement and relocation of things), andsecurity management. Sophisticated management systems may be location- and network-aware —for example, only applying large software updates over low-cost high-speed networks such as Wi-Fi. Some device management systems also collect data from things to store in the cloud. The IoTalso brings new problems of scale to the management task. Tools must be capable of managingand monitoring thousands and perhaps even millions of devices. The precise management needswill vary depending on the types of "things." Devices connected by cellular networks will often use aplatform provided by the cellular network operator for monitoring and control functions related tocellular operations, and some of these platforms provide additional management functions such asfirmware updates.

When? Management of complex IoT devices built using high-level operating systems such asAndroid is relatively straightforward because it can exploit platforms derived from related tasks suchas mobile device management (MDM). Relatively mature platforms also exist in a few vertical areas;for example, BlackBerry provides an IoT platform for the automotive industry that includes a widerange of secure management functions integrated with QNX. However, secure management of very


large numbers of simple "things" is much less mature, as are tools such as IoT platforms that mayprovide some device management features.

Who will be impacted? Any organization delivering large numbers of nontrivial connected "things"that need any form of postdeployment control or management.

Cautions: Many management platforms will be tactical decisions, so it may be necessary to changeplatforms during the life of long-lived things. Vendors selling tools derived from MDM areinexperienced in IoT and may not provide appropriate features or pricing models, and they may besubject to disruption as the mobile management market will consolidate significantly. The use ofspecific platforms mandated by the operator may be required for cellular-connected things.

Low-Power, Short-Range IoT Networks

What is it, and why is it important? Selecting a wireless network for an IoT device involvesbalancing many conflicting requirements, such as range, battery life, bandwidth, density (number ofconnected devices in an area), endpoint cost and operational cost. One important cluster of IoTnetworking technologies is focused on short range (tens to hundreds of meters), long battery life(years), relatively low bandwidth, low endpoint cost and medium density (hundreds of adjacentdevices). For example, such networks will be essential in the smart home and smart office.Topologies include point-to-point, star and mesh networks (see Note 3). Some networks extendbeyond basic communications to implement higher levels of the IoT stack, such as authenticationand security. Network selection depends not only on technical factors but also on commercialissues such as what types of devices the network must talk to; for example, because few mobilephones support it, ZigBee is unsuitable for "things" that must talk directly to a smartphone.

In terms of unit shipments, low-power, short-range networks will dominate wireless IoT connectivitythrough 2025, far outnumbering connections using wide-area IoT networks. However, thecommercial and technical trade-offs discussed earlier mean that many solutions will coexist, with nosingle dominant winner and clusters emerging around certain technologies, applications and vendorecosystems. Current technologies include ZigBee, Bluetooth, Zwave/G.9959, Thread, Ant and Wi-Fiplus point-to-point systems on a range of industrial, scientific and medical (ISM) bands. Morespecialized technologies include examples such as EnOcean, which is used to implement battery-free remote switches by using energy harvested from the action of pressing the switch.

When? More than 10 technologies of this type exist today, and new technologies and variants ofexisting ones will emerge over the next five years. Examples of technologies likely to gain traction inthe future include mesh Bluetooth, which could be important in the smart home, and Thread, whichis sponsored by Google. Academic researchers are also working on novel ultra-low-power solutionsbased on ideas such as modulating ambient signals from nearby sources such as Wi-Fi.

Who will be impacted? Any organization designing wireless "things" for home, office or personaluse.

Cautions: Because no technology or ecosystem will win the battle for the smart home or office,many environments will likely require gateways to convert between wireless protocols and devices.


The IoT implies many more objects using wireless, which could create noise and interferenceissues, so network designers must consider the impact of new wireless things on existing servicessuch as Wi-Fi.

IoT Processors

What is it, and why is it important? The processors and architectures used by IoT devices definemany of their capabilities, such as whether they are capable of strong security and encryption,power consumption, whether they are sophisticated enough to support an operating system,updatable firmware, and embedded device management agents. As with all hardware design, thereare complex trade-offs between features, hardware cost, software cost, software upgradability andso on. In the IoT space, the choice of processors is often driven by which hardware and software iseasily available to innovators, hence devices such as Arduino and Raspberry Pi have spawnedmany new products because the hardware and software tools were widely available at a very lowcost to makers and experimenters.

Gartner predicts that low-end, 8-bit microcontrollers will dominate the IoT through 2019, whichimplies that many IoT devices will be extremely simple and incapable of running an operatingsystem or performing sophisticated functions such as encryption unless built in as a hardwarefeature of the chip. Shipments of 32-bit microcontrollers will overtake the 8-bit devices by 2020; 16-bit processors will never become dominant.

Most enterprises won't design their own hardware devices, but will purchase them from externalsources or commission specialist hardware designers to create them. However, system architectsand security specialists must understand enough about IoT hardware to specify technologies thatwill satisfy their current and future business goals.

When? A very wide range of IoT processors is already available. Through 2020, the IoT will slowlybecome smarter as the use of 32-bit devices grows, enabling things to become more sophisticatedand manageable.

Who will be impacted? System architects and staff responsible for security must understand whatprocessor choices mean in terms of features, upgradability, cost and security.

Cautions: Understanding the implications of processor choices demands fairly deep technicalskills. The IoT processor market is also a battleground where companies such as Intel are trying todisplace established architectures such as ARM.

IoT Operating Systems

What is it, and why is it important? Traditional operating systems (OSs) such as Windows and iOSwere not designed for IoT applications. They consume too much power, need fast processors, andin some cases, lack features such as guaranteed real-time response. They also have too large amemory footprint for small devices and may not support the chips that IoT developers use. Evenplatforms like Android, which has emerged from the mobile phone domain (which is a type of IoTapplication), are too complex and power-hungry for many embedded systems. Consequently, a


wide range of IoT-specific operating systems have been developed to suit many different hardwarefootprints and feature needs. Several examples are illustrated in Table 1.

Table 1. Selected IoT Operating Systems

Examples of IoT Operating Systems Approximate Memory Footprint

None — device uses bare hardware or some form of support library None

Minimal embedded OS — e.g., Contiki, TinyOS, RIOT, Yottos 5KB to 50KB

Small footprint embedded OS — e.g., VxWorks, cut-down Linux 750KB to 1MB

Google Brillo 16MB

Full Linux with GUI, Android 0.5GB to 1GB

Source: Gartner (January 2016)

The minimal embedded IoT OSs are often very different from traditional platforms; many haveevent-driven programming models and some don't support modern high-level programminglanguages and features like multithreading and dynamic memory allocation. More modern minimalOS platforms such as Yottos and RIOT attempt to provide a more programmer-friendly environmentwithout excessively compromising power consumption and memory footprint. In less constrainedIoT devices with around 1 gigabyte (GB) of memory, IoT operating systems become more likeconventional OSs, although they may still lack some features such as a sophisticated GUI.Organizations creating embedded IoT software will need to master the operating systems,programming models and development tools for small-footprint devices. Designers mustunderstand the business trade-offs implied by OS selection; for example, a simple OS might implyless expensive hardware but more complex software development and could limit futureupgradability.

When? A wide range of IoT operating systems is available. New embedded OS platforms are likelyto emerge over the next five years, often from academic and open-source backgrounds.

Who will be impacted? Staff responsible for selecting, designing or developing embedded IoTapplications.

Cautions: Embedded OS systems created by open-source projects are still relatively nichetechnologies, and skills and support may be scarce.

Low-Power Wide-Area Networks

What is it, and why is it important? Traditional cellular networks don't deliver a good combinationof technical features and operational cost for those IoT applications that need wide-area coveragecombined with relatively low bandwidth, good battery life, low hardware and operating cost, andhigh connection density. The long-term goal of a wide-area IoT network is to deliver data rates from


hundreds of bits per second (bps) to tens of kilobits per second (kbps) with nationwide coverage, abattery life of around 10 years, endpoint hardware cost of around $5, and support for hundreds ofthousands of devices connected to a base station or its equivalent. Such technologies aresometimes known as low-power wide-area (LPWA) networks or LPWANs.

A range of LPWA networking technologies have emerged to satisfy some of these needs; examplesinclude Sigfox, Ingenu Random Phase Multiple Access (RPMA), LoRa, Long Term Evolution formachine-type communications (LTE-M), and more recently, a new cellular IoT standard,NarrowBand IoT (NB-IoT). A new variant of Wi-Fi — 802.11ah — operating at below 1 gigahertz(GHz) offers a range of around 1 km and may be suitable for some types of wide-area IoTapplications. Some of these technologies can be deployed as private IoT networks; others are runby carriers.

When? Proprietary technologies such as LoRa, Sigfox and Ingenu RPMA are available now, and802.11ah equipment will likely become available in 2016. Cellular IoT trials are being conducted withprestandard solutions, although standards likely won't be locked down until mid-2016, and widelyavailable, proven endpoint hardware isn't likely until 2017 to 2018.

Who will be impacted? Any organization deploying low-bandwidth IoT devices over wide areasshould explore LPWA networks. Example applications include smart cities, utility meters,environmental monitoring, equipment tracking and telemetry.

Cautions: The available technologies vary by country and sometimes even by city. There is as yetno established global standard with roaming support. Even when cellular IoT standards mature, thewide range of frequencies used by cellular networks may make international roaming difficult. ManyLPWAs are proprietary, and some operate in unmanaged public spectrum, so they can't guaranteequality of service. The long-term commercial success of the proprietary technologies is uncertain.Cellular IoT is likely to dominate in the long term, but it is not yet available, which makes strategicchoices challenging. Many of these technologies haven't achieved sufficient economies of scale todeliver on the promise of very low costs. Organizations developing long-lived "things" may need tomake tactical networking decisions and modularize their hardware so that it can be replaced whenbetter technologies are available.

Event Stream Processing

What is it, and why is it important? Some IoT applications will generate extremely high data ratesthat must be analyzed in real time (see the IoT Analytics section). Systems creating tens ofthousands of events per second are common, and millions of events per second can occur in sometelecom and telemetry situations. Traditional IT architectures that store and subsequently processdata don't have the necessary performance to deliver real-time analysis of such data streams, andin any case, there may be too much data to store in its raw form. To address such requirements,distributed stream computing platforms (DSCPs) have emerged. They typically use parallelarchitectures to process very high-rate data streams to perform tasks such as real-time analyticsand pattern identification. Examples include Apache Storm, Apache Spark, Google Cloud Dataflowand IBM InfoSphere Streams.

When? Several platforms are available now, although all are immature.


Who will be impacted? This will impact any organization whose IoT applications create extremelyhigh-rate data streams that must be processed in real time — for example, telecommunications oroil industry telemetry.

Cautions: DSCPs are a new and immature product category. No vendor meets all needs, there arefew standards, and applications cannot easily be ported from one system to another. Skilleddevelopers are scarce.

IoT Platforms

What is it, and why is it important? IoT platforms bundle many of the infrastructure components ofan IoT system into a single product. The services provided by such platforms fall into three maincategories: low-level device control and operations such as communications, device monitoring andmanagement, security, firmware updates; IoT data acquisition, transformation and management (forexample, polling devices and storing data in the cloud); and IoT application development, includingevent-driven logic, application programming, visualization, analytics and adapters to connect toenterprise systems from vendors such as SAP. There are many such platforms (although not all ofthem provide all of these features). Example vendors include CubeNube, PTC, Eurotech, GE, BoschSoftware Innovations and Axiros.

When? Many IoT platforms are currently available from a wide range of vendors, including startupsand established companies. The market will be very volatile for several years because new entrantswill continue to emerge and established companies will buy up products to build broader IoTproduct portfolios.

Who will be impacted? Organizations developing complex IoT systems should consider suchplatforms to jump-start projects because they deliver a wide range of necessary features.

Cautions: This is a very immature market and considerable technical and commercial volatility is tobe expected. All such products are very proprietary, and migration from one to another is notsimple. Many IoT platforms are expensive, with significant upfront license fees plus costs of up to afew dollars per "thing" per month, so they are unlikely to be viable for applications involving low-cost things or where things don't generate a recurring revenue stream.

IoT Standards and Ecosystems

What is it, and why is it important? Although ecosystems and standards aren't preciselytechnologies, most eventually materialize as APIs called by IoT systems. Standards and theirassociated APIs will be essential because IoT devices will need to interoperate and communicate,and many IoT business models will rely on sharing data between multiple devices andorganizations. For example, the electric company might offer you a better price for power if theirsmart meter is allowed to control your washing machine. Many IoT ecosystems will emerge, andcommercial and technical battles between these ecosystems will dominate areas such as the smarthome, the smart city and healthcare. Many different types of IoT standards and ecosystems alreadyexist. At the lower levels of the IoT stack, examples include communications protocols such as MQTelemetry Transport (MQTT), Extensible Messaging and Presence Protocol (XMPP), Data


Distribution Service (DDS) and Constrained Application Protocol (CoAP). As we move up the stack,APIs and ecosystems such as Apple's Homekit and Google's Nest will compete to own the smarthome. Other examples of standards groups and alliances include the AllSeen Alliance, the OpenFogConsortium, the Thread Group, and the Open Interconnect Consortium.

There are too many IoT "standards," and in many cases, the concerns and technologies of thesegroups overlap or compete. Some groups have specific agendas, such as promoting their favoredtechnologies or architectures. Many vendors hedge their bets by joining multiple groups, so theircommitments to specific standards are rather tactical. There are also areas of IoT where standardsare immature or incomplete, for example IoT security.

When? In 2016, there are many IoT standards and alliances; many overlap, most are immature, andin areas where there is considerable commercial potential such as the smart home, it's likely thatthe battles between standards and ecosystems will persist for three to five years. New standardswill likely emerge in areas such as IoT security.

Who will be impacted? This will impact any organization designing or developing IoT solutions,especially job roles such as CISOs and architects who are likely to be selecting standards.

Cautions: Organizations creating products may have to develop variants to support multiplestandards or ecosystems and be prepared to update products during their life span as thestandards evolve and new standards and related APIs emerge. In some domains, such as the smarthome, it's likely that converters and adapters will be required to bridge between competingecosystems and protocols.

Additional research contribution and review was provided by: Leif-Olof Wallin, Ian Keene, EarlPerkins, Timothy Zimmerman, Dionisio Zumerle and Mike Gotta.

Gartner Recommended ReadingSome documents may not be available as part of your current Gartner subscription.

"Secure Embedded Software and Systems in the Internet of Things"

"Three Best Practices for Internet of Things Analytics"

"Market Guide for IoT Platforms"

"Move to New Technology for Analytics on Fast Big Data"

"Cool Vendors in the Internet of Things 2015"

"Hype Cycle for the Internet of Things, 2015"

"How to Manage All Your IoT Endpoints"


Evidence

Information sources used in creating this research include Gartner market forecasts; discussionswith clients, vendors and colleagues; academic papers; and reviews of crowdsourcing sites such asKickstarter.

Note 1 Examples of IoT Security Issues

Recent examples of IoT-related security flaws in 2015:

■ Hackers demonstrated their ability to take control of a Jeep remotely, killing the engine anddisabling the brakes.

■ A Wi-Fi-enabled kettle was shown to store unencrypted network security keys that could beextracted by hackers.

■ A wearable fitness band was infected by malware that was then able to infect the smartphone,which was used to unload data from the band.

■ Man-in-the-middle attacks were demonstrated on industrial control systems such as those usedin power plants.

■ Smart TVs were discovered to be transmitting audio and video information to a third party evenwhen they were supposedly powered off.

Note 2 New Approaches to IoT Security

One example of one new approach to IoT security is physically unclonable functions (PUFs). Theseare functions whose response is based on physical properties of the electronics and varies betweenchips. So a PUF can be used to provide a hard-to-counterfeit response to challenges toauthenticate an object.

Note 3 Mesh Networks

Mesh networks are ad hoc networks formed by dynamic meshes of peer nodes. Routing andcontrol is distributed, and messages typically use a multihop strategy, jumping between nodes toreach their destination.


GARTNER HEADQUARTERS

Corporate Headquarters56 Top Gallant RoadStamford, CT 06902-7700USA+1 203 964 0096

Regional HeadquartersAUSTRALIABRAZILJAPANUNITED KINGDOM

For a complete list of worldwide locations,visit http://www.gartner.com/technology/about.jsp

© 2016 Gartner, Inc. and/or its affiliates. All rights reserved. Gartner is a registered trademark of Gartner, Inc. or its affiliates. Thispublication may not be reproduced or distributed in any form without Gartner’s prior written permission. If you are authorized to accessthis publication, your use of it is subject to the Usage Guidelines for Gartner Services posted on gartner.com. The information containedin this publication has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy,completeness or adequacy of such information and shall have no liability for errors, omissions or inadequacies in such information. Thispublication consists of the opinions of Gartner’s research organization and should not be construed as statements of fact. The opinionsexpressed herein are subject to change without notice. Although Gartner research may include a discussion of related legal issues,Gartner does not provide legal advice or services and its research should not be construed or used as such. Gartner is a public company,and its shareholders may include firms and funds that have financial interests in entities covered in Gartner research. Gartner’s Board ofDirectors may include senior managers of these firms or funds. Gartner research is produced independently by its research organizationwithout input or influence from these firms, funds or their managers. For further information on the independence and integrity of Gartnerresearch, see “Guiding Principles on Independence and Objectivity.”


http://www.gartner.com/technology/about.jsp

http://www.gartner.com/technology/about/policies/usage_guidelines.jsp

http://www.gartner.com/technology/about/ombudsman/omb_guide2.jsp

top 10 iot technologies for 2017 and 2018 -...

Documents